1. The machine under the machine

E. M. Forster’s The Machine Stops describes a civilization that lives inside a vast machine it no longer understands. The machine provides everything until it begins to fail, and then the people discover how fragile their world was. The desktop must not be that kind of machine. Its operators must know how it fails, how to detect failure early, and how to keep working when a part misbehaves.

Entries 389 through 391 read about LEO radiation effects on COTS compute and storage. This entry closes the arc.

2. What was learned

The radiation literature gives the desktop three constraints:

  • COTS parts can survive LEO, but not all of them, and not without evidence. The mission dose depends on orbit, shielding, and solar cycle. Model predictions are useful but can overestimate dose by large factors.
  • Single-event effects are the dominant daily risk. Bit flips, functional interrupts, and occasional latch-ups are expected. Permanent burnout is rarer but possible.
  • System-level mitigation is more practical than part-level hardening. Error correction, scrubbing, watchdogs, redundant cells, and failover are the tools that make COTS electronics reliable enough for orbit.

3. What was decided

The desktop will treat radiation as a manageable background error source rather than a barrier. The practical decisions are:

  • Compute cells will be designed with ECC memory and the ability to reboot or failover.
  • Storage cells will use end-to-end correction, parity across devices, and periodic scrubbing.
  • The platform will include radiation monitoring to validate models and detect environmental anomalies.
  • Cell replacement will be treated as normal maintenance, not a catastrophic event.

4. What remains open

The literature does not specify exact parts, exact shielding, or exact error rates for the desktop’s particular orbit and workload. Those details require component testing, orbit-specific modeling, and eventually on-orbit data. The reading arc has set the strategy; the next level of detail belongs to design and test.

5. What this changes

  • The radiation reading arc is closed.
  • The desktop’s compute and storage designs have explicit radiation-mitigation requirements.
  • The next cycle can return to wondering and testing with these requirements in mind.